SAN Health Utility

Reply
Occasional Contributor
Posts: 19
Registered: ‎08-25-2009

latency and I/O timeouts

Hi there,

I'm using b300 switches. I meet latency issues in my fabric. Brocade provides SAN Health, availables only with MS Windows. Does San Health will help me to find where latency issues occures ?

Regards.

Frequent Contributor
Posts: 140
Registered: ‎02-27-2008

Re: latency and I/O timeouts

Hi fc4brocade

I would use the CLI interface on each of your switches to probe them , and see what's going on.

"porterrshow" is a pretty handy command to run to give you a look at whats going on.


David

Occasional Contributor
Posts: 19
Registered: ‎08-25-2009

Re: latency and I/O timeouts

Thank you for your answer.

I know porterrshow command. And this one shows me that many ports have errors. But i don't know the reason.

Frequent Contributor
Posts: 140
Registered: ‎02-27-2008

Re: latency and I/O timeouts

Hi fc4brocade

Under which field are you seeing the errors ? If your seeing any under

enc in

crc err

Then I would start by changing SFP's or the fiber cables.

Good luck

David

Occasional Contributor
Posts: 19
Registered: ‎08-25-2009

Re: latency and I/O timeouts

Thank you for your answer.

I have errors in the fields er_enc_out and er_bad_os. What i understood is that this kind of error if often liked with hardware issues. I changed all my SFP. But it is yet the same situation. I don't have enougth fibers to change all my fibers.


But I am doubtful. All my used ports have errors. I can't think that all my ports are damaged.


Your opinion ?

Super Contributor
Posts: 260
Registered: ‎04-09-2008

Re: latency and I/O timeouts

Post switchshow and porterrshow of 2 switches in the same fabric - one thats affected and pick another one thats connected to it.

Execute portstatsclear on both switches before the first collection.

Collect o/p in a text file with a 2hr gap during peak hours.

Looks like the source of the error is not on the switch you are probing, If I remember right, enc_out's can propagate through the fabric.

The only way to isolate the problem is to observe delta of each port error and try to figure out the problem. Also look at fabriclog -s on all switches and check if there are flapping ports, if yes stop the cause, can help in isolation of problems.

Join the Community

Get quick and easy access to valuable resource designed to help you manage your Brocade Network.

vADC is now Pulse Secure
Download FREE NVMe eBook